21226
J. Phys. Chem. C 2010, 114, 21226–21233
Controlled Growth of Sub-10 nm Gold Nanoparticles Using Carbon Monoxide Reductant Lori A. Pretzer,† Quang X. Nguyen,‡ and Michael S. Wong*,†,§ Department of Chemistry, Department of Electrical and Computer Engineering, and Department of Chemical and Biomolecular Engineering, Rice UniVersity, 6100 Main Street, Houston, Texas 77005-1892, United States ReceiVed: August 22, 2010; ReVised Manuscript ReceiVed: October 17, 2010
There is a need to develop aqueous-phase synthesis methods of sub-10-nm Au nanoparticles (NPs), given their many exciting possibilities in catalysis, sensing, biomedical, and other water-based applications. Synthesizing Au NPs by reducing Au salt onto preformed NPs as seeds is a useful approach because final particle size can be finely predicted and controlled, though few studies have reported successful synthesis of Au NPs in the 1 and 10 nm size range. Here we report that water-suspended Au particles with a diameter ∼2.8 nm can be grown as large as ∼12 nm with sub-nanometer control, as verified through detailed ultraviolet-visible spectroscopy, small-angle X-ray scattering, and transmission electron microscopy measurements. With carbon monoxide as the reducing agent, this seeded-growth method results in colloidally stable Au sols. The reaction mechanism most likely involves the catalyzed oxidation of CO into CO2 accompanied by electron transfer to the gold hydroxide-chloride ionic species through the growing particle. Introduction 1
Au nanoparticles (NPs) have been synthesized for centuries. They are made by physically breaking bulk Au and by reducing molecular or ionic Au precursors to Au0 atoms,1-5 with the latter method leading to more monodisperse NPs.2,6 The general approaches to the liquid-phase synthesis of monodisperse Au NPs by reducing molecular or ionic Au precursors involve homogeneous nucleation and heterogeneous nucleation. Well controlled particle size distributions are readily achieved if nucleation (i.e., formation of ultrasmall particles or nuclei consisting of a few Au atoms) is relatively slow, if subsequent particle growth is fast, and if these processes are sequential and separated from one another in time.7-9 For homogeneous nucleation routes, this is accomplished by nucleating and growing NPs in one pot at specific temperatures, reactant and stabilizing agent concentrations, and synthesis time.1 For heterogeneous nucleation (also called seeded-growth) routes, monodisperse Au NPs are produced by carefully reducing gold salt onto smaller, monodisperse NPs (commonly referred to as seeds).10,11 Post-synthesis approaches to narrow a particle size distribution include centrifugation,12,13 electrophoresis,14,15 etching,16,17 chromatography,18,19 and “digestive ripening”, in which alkanethiol-coated Au NPs with a broad size distribution are refluxed in toluene.20,21 The controlled synthesis of Au NPs in the 1-10 nm range is desirable, since the most significant particle size effects appear in this range,1,22-25 leading to unique optical, electronic, and chemical properties potentially useful in catalytic, sensing, electronic, and medical applications.26-30 Sub-10 nm Au NPs (hereafter referred to as “small” NPs) are often synthesized through homogeneous nucleation methods, such as the Brust method, in which chloroauric acid, HAuCl4, is reduced by sodium borohydride in the presence of an alkanethiol in a nonpolar organic solvent. The thiols covalently bind to the NP * Author to whom correspondence should be addressed. E-mail:
[email protected]. † Department of Chemistry. ‡ Department of Electrical and Computer Engineering. § Department of Chemical and Biomolecular Engineering.
as surface ligands through the sulfur head, and the hydrocarbon tails extend from the surface to sterically stabilize the suspended NPs.1,23,30,31 In addition to their hydrophobic form, sub-10 nm Au NPs can be synthesized in water, using amino acids,32 thiolcontaining molecules,15,33-35 polymers,26,29,36 phosphonium ligands,37,38 polyphenols,39 dendrimers,24,40 and citrate41,42 as surface ligands. A report recently indicated that monodisperse and stable NPs (d ) 3.2-5.2 nm) can be synthesized through the reduction of gold chloride using sodium borohydride to result in sols that were colloidally stable for approximately one year without requiring stabilizing agents.43 Several groups demonstrated that sub-10 nm colloidal Au NPs can be catalytically active,44-47 and others conducted catalytic studies using such small Au NPs modified with a second metal and/or deposited on a support.1,48-53 The seeded-growth approach has been used to prepare Au NPs in various solvents,28,54 with a limited number of reports on aqueous-based synthesis of Au NPs in the 1-10 nm range. Jana et al. reported in 2001 that ∼3.5 nm Au seeds formed by reducing HAuCl4 with ascorbic acid in the presence of cetyltrimethylammonium bromide (CTAB) can grow into Au NPs in the 5-40 nm size range.23 Tsunoyama et al. later reported the use of ∼1.3 nm Au seeds stabilized by poly(N-vinyl-2pyrrolidone) (PVP) to grow NPs with diameters in the 2-10 nm range, by reducing AuCl4- with Na2SO3 and adding excess PVP.26 The synthesis took ∼3 h and required freeze-pump-thaw cycles, dialysis, and a N2 atmosphere. These NPs were found to be catalytically active for the aerobic oxidation of benzylic alcohols. Henglein and co-workers alternatively used γ-radiation to reduce KAu(CN)2 onto ∼2 nm Au NPs to grow NPs in the 10 nm size range.55,56 The irradiation step was conducted in a nitrous oxide atmosphere for 3 h, and ion-exchange was required post-synthesis to remove excess cyanide ions. In this work, we report a new water-based synthesis method for small Au NPs with tunable particle sizes in a relatively broad range (3-12 nm), using a seeded-growth approach. Gold hydroxide-chloride ions are reduced onto Au seeds, using carbon monoxide gas (CO) as the reducing agent. While not
10.1021/jp107945d 2010 American Chemical Society Published on Web 11/16/2010
Controlled Growth of Gold Nanoparticles commonly used as a reducing agent in the preparation of NPs57 (e.g., Au,58-61 Pd,62 and Pt63 NPs), CO is used to prepare submicrometer-sized Au-coated silica particles called Au nanoshells.64-67 We find that our synthesis method can be carried out at room temperature and atmospheric pressure to generate Au sols that are colloidally stable for at least 1 month, even though no additional surface ligands are used. Ultraviolet-visible absorbance spectroscopy and small-angle X-ray scattering of the sols and transmission electron microscopy of the dried NPs were used to determine particle diameters and size distributions, which were consistent with values predicted from magic cluster modeling. We propose a formation mechanism in which the Au seeds catalyze the oxidation of CO into CO2 and, in the process, Au hydroxide-chloride ions reduce and deposit onto the seed surface as Au0 atoms.
J. Phys. Chem. C, Vol. 114, No. 49, 2010 21227 TABLE 1: Molar Ratio of Au Salt Ions to Au NP Seeds Prior to CO Treatment Au salt Au salt/Au NP sample no. concentration (mM) molar ratio 1 2 3 4 5 6 7 8 9 10
0.04 0.07 0.11 0.15 0.19 0.22 0.26 0.30 0.34 0.37
1415 2829 4243 5658 7072 8486 9901 11315 12729 14144
total Au atom/Au NP molar ratio (ntot)a 1976 3390 4804 6219 7633 9047 10462 11876 13290 14705
a Calculated total number of Au atoms per particle after CO treatment.
Experimental Methods Materials. All chemicals were used as received unless otherwise noted. Tetrakis(hydroxymethyl)phosphonium chloride (THPC, 80% aqueous solution by weight), gold(III) chloride trihydrate (HAuCl4 · 3H2O, 99%), and anhydrous potassium carbonate (K2CO3, ACS reagent) were obtained from SigmaAldrich. Sodium hydroxide solution (NaOH, 1 M) was purchased from Fisher Scientific. Carbon monoxide gas (99%) was obtained from Specialty Chemical Products. Deionized water from a Barnstead NANOpure Diamond system (resistivity >18 MΩ/cm) was used for all experiments. Synthesis of Au Seeds. The method used to synthesize the Au seeds has been reported elsewhere.37,65 Briefly, a 1 wt % (∼25 mM) chloroauric acid (HAuCl4) solution was made by diluting 2.7 mL of a 127 mM solution (∼5 g of HAuCl4 · 3H2O in 95 mL of H2O) with 10.8 mL of H2O. 1.2 mL of a 1 M NaOH solution and 4 mL of a 1.2 mM THPC solution (0.4 mL of 80% THPC solution diluted in 33 mL of H2O) were added to 180 mL of water with vigorous stirring. After 5 min, 6.75 mL of the 1 wt % HAuCl4 solution was quickly added and the resulting liquid immediately turned brown, indicating that NPs were formed. Stirring was continued for another 10 min after the noted color change. The entire synthesis was carried out at room temperature (23 °C). The resulting sol was kept at room temperature for 3 weeks prior to any use; it was subsequently kept in a refrigerator at 4 °C for storage. The concentration of NPs was calculated to be ∼9.6 × 1017 NP/L ) ∼0.002 mM, assuming all the Au salt reduced to metal and modeling the 2.8 nm Au NPs as a ∼3.0 nm “magic cluster” containing 561 Au atoms.68 Growth of Au Seeds. A Au salt solution (pH ∼ 7.5) was prepared by dissolving 50 mg of K2CO3 in 203 mL of a 0.38 mM HAuCl4 solution (3 mL of 1 wt % HAuCl4 diluted with 200 mL of H2O), and aging for at least 24 h before use.65 Ten samples with 0.05 mL of the seed sol and different volumes (0.30-3.0 mL) of the Au salt solution were prepared and diluted with water such that the total volume of each sample was 3.05 mL and the total Au3+ concentration increased by ∼0.04 mM with increasing sample number. Deaerated water was not necessary for successful growth of Au seeds. The final Au NP concentration of each sample was estimated to be ∼2.0 × 1016 NP/L or ∼0.03 µM; the salt:seed molar ratios are listed in Table 1. Subsequently, CO was bubbled into each sample at an average flow rate of 3.3 mL/s for 30 s. The gas was bubbled into each sample for 30 s again, 1 h later, to ensure complete reduction of the Au salt. The samples were analyzed no earlier than 1 h after the second CO bubbling step. Each sample was synthesized at least three times, to ensure reproducibility of synthesis and
to assess any batch-to-batch variations. Caution: CO is toxic and all experiments should be performed in a Ventilated hood. Control Experiments. To determine if particle growth could occur without the Au seeds, ten samples were prepared in the same manner as indicated earlier, but water (0.05 mL) was used instead of the Au seed sol. To determine if particle growth could occur without the CO, ten additional control samples were prepared as indicated earlier, but CO bubbling steps were not carried out. Characterization of Au Sols. UV-vis absorption spectra (UV-vis) of the samples were collected from 400 to 900 nm using a Shimadzu UV-2401 PC spectrophotometer; polystyrene cuvettes (Sarstedt AG & Co.) with a path length of 1 cm were used. Transmission electron microscopy (TEM) images were collected using a JEOL 2010 transmission electron microscope operating at 100 kV. Au NPs were deposited onto 200-mesh carbon/Formvar grids by evaporating ∼8 µL of a sample at room temperature. The number-average size distributions of 400+ particles were determined for each sample using ImageJ Software. For small-angle X-ray scattering (SAXS), sample sols were sealed in borosilicate glass capillaries (1 mm diameter and 10 µm wall thickness) and analyzed using a Riguku SmartLab X-ray diffractometer (Cu KR radiation, λ ) 1.54 Å). All measurements were taken in transmission mode at maximum power (40 kV and 44 mA). Number-average particle sizes and distributions were determined by simulating the particle scattering profiles from 2θ values of 0.2-4.0° using Nanosolver software (Rigaku). The pH of NP suspensions was measured at room temperature using a Phoenix Co. 5533501 pH electrode. Results and Discussion Size Analysis of Au Sols. UV-vis spectroscopy can provide size information about the Au NPs while they are in suspension. The UV-vis spectra for each sample showed a systematic increase in absorbance, correlated to the amount of Au salt available for chemical reduction (Figure 1). Three separate batches were prepared and analyzed for each sample. The nearly overlapping absorbance spectra for each sample indicated the seeded-growth procedure was robust and repeatable. The sols produced did not require additional stabilizing agents and were stable for at least one month. By accounting for the particle size effect on the conduction electron mean free path in gold, one can accurately determine average Au NP size using surface plasmon resonance (SPR) peak position and absorbance. For Au NPs with diameters in the 35-110 nm range, Haiss et al.69 developed the following
21228
J. Phys. Chem. C, Vol. 114, No. 49, 2010
Pretzer et al.
Figure 1. (a) UV-vis absorbance spectra and (b) corresponding photograph of Au sol samples after CO treatment. Each sample was prepared three times, and an absorbance spectra of each batch was collected.
TABLE 2: Sample Particle Diameters (nm) Determined Through Different Methods sample no. seed particles 1 2 3 4 5 6 7 8 9 10
UV-visa b
2.4 (4.0%) 4.0 (5.0%)b 4.7 (2.0%)b 5.3 (3.0%) 5.9 (1.0%) 6.7 (7.0%) 7.0 (1.0%) 7.4 (1.0%) 7.9 (1.0%) 8.3 (1.0%) 8.6 (1.0%)
predictedc
TEM meand
TEM mode
SAXS meand
SAXS mode
3.0 4.5 5.4 6.1 6.6 7.1 7.5 7.9 8.2 8.6 8.9
3.0 ( 0.9
2.8
7.3 ( 1.6
7.1
11.9 ( 4.6
10.1
3.6 ( 1.3 5.7 ( 2.0 6.1 ( 2.2 6.8 ( 2.5 7.2 ( 2.6 7.7 ( 2.8e 8.5 ( 3.1 8.6 ( 3.2 9.2 ( 3.4 9.4 ( 3.4 9.7 ( 3.6e
3.1 4.9 5.3 5.8 6.2 6.7 7.3 7.5 7.9 8.1 8.3
e
a
Sizes calculated from absorbance spectra using eq 2. RSDs (%) represent experimental error between three separate syntheses. b Values outside the size range for eq 2. c Sizes predicted using a power fit to the magic cluster model, eq 4. d Size distributions represent one standard deviation. e Particle sizes and distributions averaged for three separate syntheses.
correlation (with an error of 3%) between experimental absorbance results for citrate-stabilized Au NPs and Mie scattering theory:
D ) ln[(λSPR - 512)/6.53]/0.0216
(1)
where D is the mean NP diameter from TEM measurements and λSPR is the SPR peak position (ranging from 525 to 585 nm). More appropriate for our study was the following size correlation (with an error of 11%) based on the relative absorbance for Au NPs in the 5-80 nm size range:
D ) exp([B1 × (Aspr /A450)] - B2)
(2)
These absorbance-derived size values were compared with particle sizes predicted using the magic cluster model. With this model,25,70,71 a Au NP is approximated as a single Au atom surrounded by closed shells of Au atoms,68 such that the total number of atoms, ntot, is exactly known for a magic cluster of ns number of closed shells: ntot ) (10ns3 + 15ns2 + 11ns + 3)/3.68 With a Au atom diameter of 0.27 nm, an analytical relationship between magic cluster diameter Dmc and ntot, can be derived (eq 3), which simplifies to a power-law equation showing a 1/3 dependence of diameter on number of atoms (eq 4):
Dmc ) 0.27 × (1 + 2ns) where ns ) (0.033f) - (3.50/f) - 0.5 and
where ASPR is the absorbance of the SPR peak, A450 is the absorbance at 450 nm, and B1 ()3.00) and B2 ()2.20) are fitted parameters using experimental data.69 Application of eq 2 provided mean NP diameters that ranged from 2.5 nm for the seed particles to 8.6 nm for the Au NPs of sample 10 (Table 2). Standard error propagation analysis based on the triplicate syntheses of each sample indicated high reproducibility, with the maximum error, or relative standard deviation, found to be 7% (one standard deviation divided by mean diameter ) relative standard deviation ) RSD). The diameters of the seed particles and NPs of samples 1 and 2 were outside the operational range of eq 2, and so these extrapolated UV-vis-derived sizes, while precise, may not be accurate.
f ) [4,050ntot + 15 × (5,145 + 72,900ntot2)1/2]1/3
(3) Dmc ≈ 0.359ntot0.334
(4)
Equation 4 matches well with the expected relationship between particle diameter and total number of atoms for fcc-packed metal atoms.25 Assuming that all of the Au salt precursor for a given sample completely reduces and deposits onto the Au seeds after CO treatment, the Au salt/Au NP molar ratio (Table 1) represents the additional atoms added to one seed particle and the Au atom/Au NP molar ratio is the total atoms per particle
Controlled Growth of Gold Nanoparticles TABLE 3: Sample Particle Concentrations Estimated from NP Diameter and Absorbance at 450 nm sample
NP concentration (×1016 NP/L)
seed particles 1 2 3 4 5 6 7 8 9 10
2.6 ( 0.7 2.2 ( 0.3 2.3 ( 0.4 2.2 ( 0.7 2.0 ( 0.5 1.7 ( 2.3 1.7 ( 0.3 1.7 ( 0.3 1.6 ( 0.7 1.5 ( 0.7 1.5 ( 2.8
after growth is complete. As shown in Table 2, the absorbancederived diameters of sample sols were in good agreement with the diameters predicted by eq 3 in which each seed particle is assumed to consist of 561 Au atoms (Dmc ) 2.97 nm). Having determined the size of the different Au NPs from their absorbance spectra, we estimated the NP concentration in NP/mL, N, in our samples using another fit developed by Haiss et al.,69
N ) (A450 × 1014)/(d2 × [-0.295 + 1.36 exp(-((d - 96.8)/78.2)2)] (5) where d is the NP diameter (nm). Application of eq 5 gave NP concentrations that ranged from 2.6 × 1013 (for Au seeds) to 1.5 × 1013 (for sample 10) NP/mL, with the apparent concentration differences being within error (Table 3). These results are also within error of the estimated concentration of Au seeds (∼2.0 × 1016 NP/L) assuming complete reduction of gold chloride to form monodisperse ∼3.0 nm Au seeds. TEM was performed on selected samples to quantify the Au particles in the dried state (Figure 2). The seed particles had a mean diameter of 3.0 nm and a mode diameter (i.e., the most frequent diameter exhibited by the particles) of 2.4 nm. The seeds were unimodal in size but were broadly distributed, with a RSD of 30%, which is close to previously reported values.37 Samples 5 and 10 showed systematic growth in diameter size, after CO treatment (Table 2). The RSD’s (22% and 39%, respectively) were variable and moderately large, which could be due to the broad size distribution of seed NPs or drying effects and electron beam damage during TEM analysis.72 The mean diameters were at least 17% larger than the mode diameters for these 3 suspensions, indicating a positive skew to the size distributions for all samples. Small-angle X-ray scattering (SAXS) was used to determine the diameter and size distribution of the NPs dispersed in water (Figure 3). A small aliquot of our samples was sealed in a glass capillary tube and irradiated with X-rays. With the particles having a higher electron density than water, the collected scattered X-ray spectra contain information about the numberaverage size and shape of the NPs, which can be extracted by comparison with simulated scattering spectra of model NP suspensions.73,74 Reliability factor (R factor) values between 1 and 10% indicate an acceptable match with lower values (
